A Framework for the Semi-Automatic Testing of Video Games

Abstract

Game environments are complex interactive systems that require extensive analysis and testing to ensure that they are at a high enough quality to be released commercially. In particular, the last build of the product needs an additional and extensive beta test carried out by people that play the game in order to establish its robustness and playability. This entails additional costs from the viewpoint of a company as it requires the hiring of play testers. In the present work we propose a general software framework that integrates Artificial Intelligence (AI) Agents and Computer Vision (CV) technologies to support the test team and help to improve and accelerate the test process. We also present a prototype shadow alias detection algorithm that illustrates the effectiveness of the framework in developing automated visual debugging technology that will ease the heavy cost of beta testing games.